text stringlengths 0 2.35k |
|---|
**Mat Ryer:** See, I think that might be a good one for an unpopular one. I think that might be unpopular. So you're saying before you import a package, read every line... |
**Egon Elbre:** Yes. |
**Mat Ryer:** What about its dependencies? |
**Egon Elbre:** Those as well, of course. |
**Roger Peppe:** And presumably, every time you update to a new version, right? You have to review all the changes, too. |
**Egon Elbre:** Yeah. I think people import packages assuming they are better than their codebases, right? And if you have such really high standards for your own base, why wouldn't you have those as well for other codebases? |
**Mat Ryer:** Yeah. Isn't it about responsibility? |
**Egon Elbre:** Yeah. |
**Mat Ryer:** Sometimes there are packages out there and they maybe have a couple of stars, they're not really used by anybody... But there are clues, there are signals to look for for packages that are sort of a bit meatier and a bit more stable. Is it our responsibility? |
**Roger Peppe:** \[52:06\] I have to say that I don't do that. What I do tend to do is when I look at a dependency, I look through the code and say, "Does this look what I would expect to find in a package that is implementing that functionality? Does it generally look good, feel about right? Is this too big for what i... |
**Mat Ryer:** Yeah. "Does it have tests?" That's a quick check... |
**Roger Peppe:** To be honest, I struggle to understand my own code, let alone other code that I'm importing... So I couldn't understand the gRPC code, the HTTP/2 code... You know, somebody's spent man years, man centuries probably working on that code, and there's no way I could meaningfully review it, I don't think. |
**Mat Ryer:** Man or woman years. They years. They years. \[laughter\] That's the modern measure of time. |
**Roger Peppe:** Sorry. |
**Mat Ryer:** Yeah. That's all right. |
**Roger Peppe:** Person years. |
**Mat Ryer:** Person years. Yeah... Yeah, go on, Egon? What do you reckon? |
**Egon Elbre:** But yeah, I think it has many benefits, if you do review them, even if you don't understand then, right? Because if you are reviewing other code, you might look at how they build things. Maybe you learn something new, right? Maybe you discover some bugs that need to be fixed, because you have different ... |
Of course, when I say you review them with the same standard, you don't worry about formatting details, or they decided to use spaces instead of tabs. Like, let them have those spaces... |
**Roger Peppe:** I do. No way can they use spaces. Just, absolutely no. No. Out. \[laughter\] |
**Egon Elbre:** And also, many of the -- there have been many attacks against code injections, and -- that you should be worried about, right? |
**Mat Ryer:** It's one of those that's hard to argue against what somebody said. It's like if someone says you should have all this extra security; it's hard to say, "No, we shouldn't." But practically... |
**Roger Peppe:** I think it's easy to argue against, because I don't think it's reasonable. I don't think it's possible for most people. I think that we should have better supply -- you know, better dependency chain assurances, honestly, that we do. |
**Mat Ryer:** But until then... |
**Egon Elbre:** I do think there is a case where you don't need to review them. If you're actually paying someone else to maintain that codebase, then you're intentionally offsetting your own responsibility to somebody else. I do consider every dependency as your own responsibility to maintain, and to fix things when t... |
**Mat Ryer:** I mean, with that rule in place, you will certainly be incentivized quite strongly not to include many dependencies. So it has that effect, too... \[laughs\] |
**Egon Elbre:** Of course, yes. Like, do you really want to read the gRPC? |
**Mat Ryer:** I do now... I'm going to print it out, and take it to bed. |
**Egon Elbre:** Maybe you can make a bed out of it already. |
**Mat Ryer:** \[laughs\] Yeah. Nice. Okay, Roger, have you got an unpopular opinion for us? |
**Roger Peppe:** Yeah. My unpopular opinion is that I think that often just working on a laptop with just a small screen - not on my big desk, with multiple monitors... Just my laptop, with my little screen, and just a keyboard, maybe just on a sofa or something, it's actually more productive than working on my big des... |
**Mat Ryer:** Yeah. So is this about like there's no distractions, and you're just in that world? |
**Roger Peppe:** I think it might be. I don't really understand it, to be honest, because it should be that I've got all these -- I can see all the things, I can do all the things, I've got the nice, rising desk, all the stuff that's absolutely optimized for me... And then I can sit on the sofa with my laptop, and sudd... |
**Mat Ryer:** What do you think of that, Egon? |
**Egon Elbre:** \[55:58\] I do think it's nice. I occasionally go to coffee shops to program there, and it does give you a bit more focus. |
**Mat Ryer:** Yeah. I also wonder, Roger, if it's because most of your time has been spent there, rather than -- if you think about in the past, you didn't have all this tech and tease monitors and this desk that moved up and down. |
**Roger Peppe:** In the past you didn't have a laptop. Laptops are a new thing, really... \[laughs\] |
**Mat Ryer:** Hm... You just had to have your monitor on your lap if you wanted to do that. |
**Roger Peppe:** It was a bit big and heavy... That big 17-inch thing. |
**Mat Ryer:** By the way, instead of those desks that go up and down - it's a bit more expensive, but you can actually just have your entire floor move, and same effect; have your desk fixed, and then you just move with the floor. |
**Roger Peppe:** It's funny, because I've got this rising desk, which technically has four saved positions in there... And I actually never use it anymore. I just stand. I started off and I'd stand and I'd sit and now I just, I might as well just get a desk of that height. |
**Mat Ryer:** Oh yeah, yeah. |
**Roger Peppe:** I never sit down. Apart from when I'm on the sofa. Maybe that's the reason. |
**Mat Ryer:** But why do you need four? You've got standing, sitting... Is there one where you're on your knees, praying, really hoping the code is going to work? |
**Roger Peppe:** My theory is that it's for you and one other person. |
**Mat Ryer:** Yeah, that's my theory, too. But for comic effect, I pretended not to know that. |
**Roger Peppe:** For comic effect...? \[laughs\] |
**Mat Ryer:** Yeah, it doesn't always work... |
**Roger Peppe:** Ha-ha-ha. |
**Mat Ryer:** Ha-ha! Thank you. \[laughter\] Let me ask you this, Roger, as well. Are you a keyboard wizard type? Do you use Vim, Emacs, that kind of thing? |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.